29 November 2019
About The Client
Frrole is a best-in-class AI platform based in Bangalore that works at the intersection of social data and consumer intelligence. They believe that the software must become more intelligent to continue delighting the customer. Frrole has mastered the social data set that can be best utilized to understand consumers at scale and to build the Google of the social web, unleashing the power of accurate and predictive consumer insight for all the software in the world, making it truly intelligent.
The Business Need
The client uses social media data to build the Google social web. They were already using Azure cloud services for its two flagship products, Scout and Deeper Sense. However, they wanted to migrate to AWS for two major reasons:
- Reduce the overall cost,
- Apply best practices to help scale.
Thus, Frrole sought an AWS expert to help them successfully migrate.
Since both applications were business-critical, the client wanted minimal downtime during the migration. Also, the amount of data to be migrated was in terabytes, and the entire migration had to be done within extremely short timelines.
After studying the working of the applications, cloud experts at Rapyder came up with an optimum solution for migrating the application from Azure to AWS. Our cloud experts stitched a solution with AWS services, and the following solutions were implemented:
- A VPN connection from AWS to Azure account was provisioned for seamless connectivity.
- The NoSQL database was extended into AWS, considering the vast amount of data to be transferred.
- The search engine of the application had terabytes of data in the clusters of AWS by extending the clusters into AWS, which ensured zero downtime.
- All other servers were created in a private subnet only to be accessed within a VPC for security with a load balancer.
- The NoSQL database and search engine cluster were deployed across availability zones for HA.
- Open VPN server was provisioned to connect directly to the instances in private servers locally.
- NAT gateway was used for all outgoing internet connections from private servers.
- Reserved instances recommendations helped in reducing the cost further after.
- Considering the timeline requested by the client, the entire migration, including the database of 5 TB, was completed within 10 days.
- After migrating the entire application to AWS, the client recognized a cost saving of almost 25% compared to the expenses paid on Azure.
- The solution provided is highly available and scalable at all application layers.